The enigma of the triangular pyramid (Q1266474)

In Euclidean space \(\mathbb{R}^3\), let \(T\) be an arbitrary tetrahedron (triangular pyramid). Denote by \(d(T)\) the distance between the centres of the insphere (radius \(r)\) and the circumsphere (radius \(R)\) of \(T\). Then the non-negative number \(g=g(T)\) defined by \(g^2=(R+r)(R-3r)\) is called the ``Durrande number for the tetrahedron \(T\)''). In 1823/24, J. B. Durrande ``proved'' (thus solving a problem posed by J. D. Gergonne in 1816) that \(d(T)=g(T)\) be valid for any tetrahedron T. However, this is not true. In this article the authors show: There is no universal, suitable for all tetrahedra, formula \(d=f(R,r)\); \(d(T)=g(T)\) is true for every regular triangular pyramid (=pyramid whose base is a regular triangle and whose altitude passes through the centre of the base), but \(d(T)<g(T)\) holds if \(T\) is a rectangular tetrahedron which is not a regular pyramid. (Remark: first concrete counter-examples to Durrande's proposition are due to G. Danielsson in 1946). From these results two questions arise: (1) Does there exist at least one tetrahedron \(T\) which is not a regular pyramid but satisfies the condition \(d(T) =g(T)\)?, and (2) Is the inequality \(d(T)\leq g(T)\) valid on the set of all tetrahedra? In both cases, computer experiments have made the answer ``yes'' highly plausible.